AnsweredAssumed Answered

AD9361 "Fast Acting" Rx Input Protection?

Question asked by RF-guy on Jul 21, 2015
Latest reply on Sep 29, 2015 by RF-guy

On Jan 22, 2014, there was a forum question titled "What is the failure mechanism of the AD9361" that asked about what kind of failure would happen if the Rx input max spec of +2.5dBm (single-ended) were exceeded.  We're facing a similar issue with trying to decide on how to best protect these Rx inputs.  However, ADI doesn't seem to supply guidance on how quickly input protection needs to react to transients or overdrive.

 

In the post mentioned above, the last ADI response from "tlili" states: "It is recommended to provide fast acting protection at Rx inputs to limit signal amplitude to the following levels: +2.5dBm (at 50ohms) single ended or +5.5dBm (at 100ohms) differential."

 

Also, after looking through ADI's responses to various questions about max Rx input levels and protection, this is the only response I see that indicates that even a slight overdrive might not damage the input circuits instantaneously.  This is good news because it is very difficult to prevent all input transients that might exceed 0.42Vpk (+2.5dBm peak).  If the chip could be instantaneously damaged, then it would be useless in a real RF environment, especially since input protection circuits usually have some small delay.

 

So, can ADI give us any guidance at all on a rough order of magnitude of overdrive time it might take to damage the input (nanoseconds, microseconds, milliseconds, seconds, etc.)?  I know this damage time is probably dependent on the amount of overdrive, so let's say it's briefly around +20dBm or so.  Additionally, is an Rx input failure caused by something heating up internally, or does it have more to do with some time of semiconductor insulator breakdown or similar?  If it's due to something heating up, then I suppose that chip temperature is also a factor?

 

This is an important issue for ADI to test or estimate and provide some customer guidance. If this damage time is instantaneous or too short, then protection circuits are impossible, and the chip is useless in most real-world environments.  Thanks for your help in this important matter.

Outcomes